• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Diffusing 1st reflections of speakers that measure great on and off-axis - instead of absorbing

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,523
Likes
3,745
Location
Princeton, Texas
can anyone present a valid response to the questions regarding the validity of the test environment? or is the general consensus that verification of the room characteristics is not relevant for such science-based listening tests?

I don't have answers to your questions about the test environment in the multi-channel room. I haven't delved into the preferred acoustics of a dedicated multi-channel room enough to have a sufficiently strong opinion to question Harman on the subject of verification. I have no problem with you questioning them though.

I do see limitations in the test environment for the single-speaker double-blind listening tests, but not really beyond the limitations Olive acknowledges in his paper.
 
Last edited:

Bjorn

Major Contributor
Audio Company
Forum Donor
Joined
Feb 22, 2017
Messages
1,286
Likes
2,562
Location
Norway
Harman's reference listening rooms are built exactly the same with same treatment at different locations in the world. I'm quite confident the absorbers are around 2"-2.5" thick.
Harmen-Tour-1.jpg


They are defintely not 4". Unless you use a multicore product like Broadsorbor, you need 6"-8" thick absorbers in order to treat very well down to the Schroeder frequency.

However, the room Harman uses for shuffle speaker testing is another. This one:
IMG_4998.jpg


Not sure what that black product is, but the absorbers aren't broadband here either and placement is weird. As a sidenote, it looks to me that the speakers aren't placed in the same position when testing. If that's true, that's very surprising but I may be fooled by the video.

Harman also has a multichannel room with speaker shuffling. They use 120 RPG Skylines diffusers and the goal was only to attenuate specular reflections below -10 dB. The Skyline diffusers don't diffuse very much as mentioned before, they are 2D diffusers meaning the will waste energy vertically and hence not a great choice for walls IMO and nothing has been done in order to reduce periodicity.
Harman Shuffler Testing Room.PNG

Harman Shuffler Testing Room #2.PNG
 

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
Harman's reference listening rooms are built exactly the same with same treatment at different locations in the world. I'm quite confident the absorbers are around 2"-2.5" thick.

They are defintely not 4". Unless you use a multicore product like Broadsorbor, you need 6"-8" thick absorbers in order to treat very well down to the Schroeder frequency.

Thanks, I hadn't seen these pictures before. Can you comment on normal incidence versus 45 degree incidence versus random-incidence absorption data and how that modifies the spectra of reflections? I've only seen the graphs in Toole's book (somewhat different in the two editions, with overlapping data in the first) provided by Peter D'Antonio. I would say that the averaged measurements at the listening positions were reassuring that the room is not overly damped at higher frequencies.

However, the room Harman uses for shuffle speaker testing is another. This one:

Not sure what that black product is, but the absorbers aren't broadband here either and placement is weird.

The multichannel listening lab construction was described as "The inner walls and ceiling of the double-wall IAC shell are made of heavy gauge steel panels separated 10 cm and filled with fiberglass. The inner surfaces are perforated with 2.34-mm openings to provide substantial sound absorption inside the room." I don't know the spacing or how that modifies the absorption. I also don't know about this specific room, so I find it hard to make any meaningful statements.

As a sidenote, it looks to me that the speakers aren't placed in the same position when testing. If that's true, that's very surprising but I may be fooled by the video.

It looks like the frame slides sideways and then the floor panels move backwards or forwards to me to allow for same site positioning.

Harman also has a multichannel room with speaker shuffling. They use 120 RPG Skylines diffusers and the goal was only to attenuate specular reflections below -10 dB. The Skyline diffusers don't diffuse very much as mentioned before, they are 2D diffusers meaning the will waste energy vertically and hence not a great choice for walls IMO and nothing has been done in order to reduce periodicity.

Right, I think they're achieving the desired goal, which is attenuation, not diffusion.

Bjorn, I asked you before to please comment on preferred psychoacoustics for multichannel listening.

Young-Ho
 

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
I read a very interesting book called "Acoustics of Small Rooms" by Mendel Kleiner and Jiri Tichy. It discussed ITU-R BS.1116 recommendations and the IEC 268-13 standard. I wonder if the Harman reference room departs from some of these in order to come closer to approximating a domestic living/listening environment, where ceiling treatments are less common but where bookshelves and other furnishings may provide some scattering effect.

They mentioned Aalborgh University Centre's Acoustics Laboratory. Here's the link: https://www.es.aau.dk/sections-labs/Signal-and-Information-Processing/Lab+Facilities/. Note some of the treatments in the pictures. Probably some users on the Internet will express outrage that they don't provide ETC information for their rooms.

Although they're discussing control rooms, the authors note "The surface treatment of rooms for surround sound monitoring must be designed very differently from rooms designed for stereo sound. The need for a somewhat longer reverberation time being needed for stereo has already been pointed out." They present three basic options, the first being highly absorptive and the third being highly diffusive (Blackbird Studio C). "The second type of design combines some reverberation with the removal of very-high-frequency reflection cues [19]. This design is shown in Figure 11.12. Patches of sound-absorptive treatment are combined with splayed wall sections that reflect high-frequency sound away from the listener so that there are no image sources visible from the listener’s position."

Screen Shot 2020-07-30 at 5.25.22 PM.png

To my ignorant eye, the Harman reference listening room doesn't seem to depart from this so radically as some users' comments might suggest, other than the absence of reflectors on the front wall. Given the knowledge that Skyline diffusors do result in some absorption, I see hemicylindrical elements being used on the side and rear walls where the "reflectors" are suggested, with Skylines also on the sidewalls in the other "reflector" positions. Given the lobing issues with polycylindrical arrays, limiting the number makes sense to me. Interesting that the authors focus on high frequency reflection cues, almost reminds about what Toole said about flutter echo in his books, which again, have suggestions mostly in line with what I see in the Harman reference listening room when it comes to multichannel.

Young-Ho
 

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
To my ignorant eye, the Harman reference listening room doesn't seem to depart from this so radically as some users' comments might suggest, other than the absence of reflectors on the front wall. Given the knowledge that Skyline diffusors do result in some absorption, I see hemicylindrical elements being used on the side and rear walls where the "reflectors" are suggested, with Skylines also on the sidewalls in the other "reflector" positions. Given the lobing issues with polycylindrical arrays, limiting the number makes sense to me. Interesting that the authors focus on high frequency reflection cues, almost reminds about what Toole said about flutter echo in his books, which again, have suggestions mostly in line with what I see in the Harman reference listening room when it comes to multichannel.

The setup/orientation of the Harman reference listening room is also more in line with the sort of audience experience suggested by a significant fraction of images of home theater setups on the Internet that I have personally encountered. Since Harman is in the business of selling products, not the purely academic pursuit of knowledge, it makes sense to me that they don't use the Blackbird Control Room C for their listening tests (interesting that the original design was subsequently modified).

Kleiner and Tichy go on to discuss home listening rooms, noting "The experience gained from control room design is useful in the design of domestic listening rooms...In domestic environments, the reverberation time at high frequencies is typically determined by tapestries, carpets, drapes, and furniture...Shelves filled with books offer substantial sound absorption because of the porosity of books...the presence of large scattering objects, localized sound absorption by bookshelves and furniture, and asymmetrically placed building details such as doors and windows."

Young-Ho
 

Kal Rubinson

Master Contributor
Industry Insider
Forum Donor
Joined
Mar 23, 2016
Messages
5,273
Likes
9,789
Location
NYC
It looks like the frame slides sideways and then the floor panels move backwards or forwards to me to allow for same site positioning.
Yes. That is what happens. BTW, even when the transition is from Speaker N to Speaker N (as for an A/B/X test), both physical actions are made although the speaker ends up the the same place. This is to prevent some canny subject from assessing what is going on behind the screen from the carriage noises.
 

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
Thanks, I hadn't seen these pictures before. Can you comment on normal incidence versus 45 degree incidence versus random-incidence absorption data and how that modifies the spectra of reflections? I've only seen the graphs in Toole's book (somewhat different in the two editions, with overlapping data in the first) provided by Peter D'Antonio. I would say that the averaged measurements at the listening positions were reassuring that the room is not overly damped at higher frequencies.

For those of you who don't have the first/second edition handy, here is one figure relevant to what I mentioned above for what Toole refers to as frequency-dependent absorption.
Screen Shot 2020-07-30 at 5.55.13 PM.png
Toole presents additional 30 degree incidence data from D'Antonio in the third edition.

It almost looks to me like fabric-covered absorption panels perform more evenly when positioned for closer to normal incidence, like, for example, on the front and rear walls. [edit: I'm thinking here of the front three channels, specifically, since they could be argued to be of primary importance]

Young-Ho
 
Last edited:

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
Yes. That is what happens. BTW, even when the transition is from Speaker N to Speaker N (as for an A/B/X test), both physical actions are made although the speaker ends up the the same place. This is to prevent some canny subject from assessing what is going on behind the screen from the carriage noises.

Thank you, Kal. Bjorn, personally, I think it can be problematic to rely too much on superficial visual inspection when looking at pictures, since parallax can create problems of perspective. In the first picture you linked above of the reference listening room, for example, look at how shallow the LSR6332 speaker looks on the left side of the picture compared to the one on the right. I could not say with any personal confidence what any of the specific measurements are in the picture, just as you could not say with any confidence what was occurring with respect to speaker position in the video. In any case, as I mentioned above, this slide https://docs.google.com/presentatio...tsaowbpiFvjs_epoBvaTv28o7w/edit#slide=id.i165 suggests to me that the overall frequency spectra has not been unduly compromised by the use of overly thin absorption elements, despite your concerns.
 

youngho

Senior Member
Joined
Apr 21, 2019
Messages
486
Likes
799
Can we be respectful, of each other and towards the research.

Let's be clear: I interpret a phrase like "stunningly absurd" to be extremely provocative and inflammatory, as well as disrespectful, and I typically consider a troll to be someone who intentionally upsets others. When a so-called apology is followed by an invitation for others to call me out for something that I didn't actually say, I don't accept it. Although I have tried to contribute positively towards this thread in a number of ways, just as I said to someone that I doubted I had anything to say of further interest to them, I think it's likely that I have little to contribute of further interest here.

":). mods, please feel free to remove my posts as applicable," or let me know, and I can delete all my posts in this thread
 

Duke

Major Contributor
Audio Company
Forum Donor
Joined
Apr 22, 2016
Messages
1,523
Likes
3,745
Location
Princeton, Texas
Not sure what that black product is, but the absorbers aren't broadband...

From Harman's paper about the Multichannel Listening Room, top of Page 5 [edit: I see that youngho already posted this]:

"The inner walls and ceiling of the double-wall IAC shell are made of heavy gauge steel panels separated 10 cm and filled with fiberglass. The inner surfaces are perforated with 2.34-mm openings to provide substantial sound absorption inside the room. The inner walls are entirely floated and separated from the outer wall of the shell by a 10 cm space to minimize mechanical and acoustic transmission of noise." [emphasis Duke's]

That sounds to me like the walls and ceiling of the room are (broadband?) absorbers in the Multi-Channel Listening Room.

Harman's reference listening rooms are built exactly the same with same treatment at different locations in the world...[emphasis Duke's]

So I take it, then, that the double-blind-speaker-shuffler room has the same (presumably broad-band absorptive) wall and ceiling construction?

From Harman's paper again, Page 7:

"The average Tm value for the MLL is about 0.23 s, which is slightly below the calculated ITU and EBU optimal value of 0.29 s. However, the curve
falls within the minimum recommended value, and is quite uniform with frequency, only rising slightly below 125 Hz."

Whatever function the fiberglass-filled perforated walls and ceilings perform, apparently a remarkable degree of broadband absorption has been achieved in that room, and presumably in any others "built exactly the same."
 
Last edited:

Bjorn

Major Contributor
Audio Company
Forum Donor
Joined
Feb 22, 2017
Messages
1,286
Likes
2,562
Location
Norway
Thanks, I hadn't seen these pictures before. Can you comment on normal incidence versus 45 degree incidence versus random-incidence absorption data and how that modifies the spectra of reflections? I've only seen the graphs in Toole's book (somewhat different in the two editions, with overlapping data in the first) provided by Peter D'Antonio. I would say that the averaged measurements at the listening positions were reassuring that the room is not overly damped at higher frequencies.
The incidence effect both how the low the product absorbs (because of travelling length) and how well higher frequencies are absorbed. How low it absorbs can be calculated if you know the air flow resitivity of the material. How well it absorbs the highs depends not only on the porous material but also on the fabric and surface of the item. A surface like the RPG Abffusor will attenuate higher frequencies better, but the product should have been made out of other materials. I've communicated with RPG in UK about this and we've discussed a new product.


The multichannel listening lab construction was described as "The inner walls and ceiling of the double-wall IAC shell are made of heavy gauge steel panels separated 10 cm and filled with fiberglass. The inner surfaces are perforated with 2.34-mm openings to provide substantial sound absorption inside the room." I don't know the spacing or how that modifies the absorption. I also don't know about this specific room, so I find it hard to make any meaningful statements.
Sound absorption from perforated panels in small rooms don't work very well and especially in regards to treat specular reflections with high gain. I've treated office rooms where perforated panels where originally used, and the room had to be fixed with better treatment.



Right, I think they're achieving the desired goal, which is attenuation, not diffusion.
I so, you use absorbers and not diffusers. One of the goals was the EBU standard with attenuation of the first 15 ms to -10 dB. That's an exeptional poor standard because the specular reflections are still highly audible.

Bjorn, I asked you before to please comment on preferred psychoacoustics for multichannel listening.
I can't say when it comes to multichannel for sure and that will also depend on the type and whether it's movies, music that's been mixed to multichannel or upmixed from two channels, etc. With movies one can generally have a "drier" room with more absortive treatment and without much diffusion because you have the spaciousness in the surround mix. Specular reflections still need to be treated with broadband treatment and you also need control of room modes and resonances. Broadband treatment combined with load of bass absorption avoids the dead feeling to a great deal. It's when bandlimited treatment are used, the room sounds dull and dead. That being said, a room with only absorption on many surfaces will have a lot less energy compared to the use of diffuse energy. Multichannel simply isn't correct or accurate, so there's any treatment that will be either. But generally I would use more absorption or combined with a hybrid product like BAD Arc compared to two channels.

For two channels and in regards to accuracy it's easier. Obviously one has to consider the room dimension, distance to surfaces and rear wall and number of listening positions. But generally you want as much late arriving lateral exponential diffuse energy as possible from high quality and broadband diffusers. An ISD gap between 20-25 ms is probably ideal. A strong termination of the ISD gap with Haas kicker can actually make it sound better, but it isn't quite as accurate to the original signal.
 

Bjorn

Major Contributor
Audio Company
Forum Donor
Joined
Feb 22, 2017
Messages
1,286
Likes
2,562
Location
Norway
From Harman's paper about the Multichannel Listening Room, top of Page 5 [edit: I see that youngho already posted this]:

"The inner walls and ceiling of the double-wall IAC shell are made of heavy gauge steel panels separated 10 cm and filled with fiberglass. The inner surfaces are perforated with 2.34-mm openings to provide substantial sound absorption inside the room. The inner walls are entirely floated and separated from the outer wall of the shell by a 10 cm space to minimize mechanical and acoustic transmission of noise." [emphasis Duke's]

That sounds to me like the walls and ceiling of the room are (broadband?) absorbers in the Multi-Channel Listening Room.
Like previously mentioned, perforated panels doesn't work very in small rooms with close boundaries and certainly not in regards to treat high gain specular reflections. Perforate panels work well in rooms with reverberation time.


So I take it, then, that the double-blind-speaker-shuffler room has the same (presumably broad-band absorptive) wall and ceiling construction?
It doesn't look like there's any ceiling treatment from the pictures. But difficult to say for sure.
IMG_4998.jpg

Harmen-Tour-1.jpg


There's is no broadband treatment of specular reflections in these rooms. It's highly bandlimited altering the spectral content. Speakers with broad dispersion is likely to be preferred in such rooms.

From Harman's paper again, Page 7:

"The average Tm value for the MLL is about 0.23 s, which is slightly below the calculated ITU and EBU optimal value of 0.29 s. However, the curve
falls within the minimum recommended value, and is quite uniform with frequency, only rising slightly below 125 Hz."

Whatever function the fiberglass-filled perforated walls and ceilings perform, apparently a remarkable degree of broadband absorption has been achieved in that room, and presumably in any others "built exactly the same."
It's based on measurements and methods that are not applicable to small rooms as has been discussed in this thread.

It's pretty clear they don't understand much about small room acoustics and trying the best to find a middle road.
In the beginning of the paper they mention the controversial of reverberation time, but still end up following it together with attenuation reflections up to 15 ms with only -10 dB.

In reviewing these various standards, a serious problem common to many is that while they define tolerances for specific acoustic parameters, they do not adequately define how the parameter is to be measured. For example, IEC is the only standard that specifies how reverberation time should be measured, even though it has been shown that RT60 can vary widely depending on the technique used. Unfortunately, this rather defeats the purpose of defining a standard in the first place! It is conceivable that one measurement method may show the room meets the standard, while another measurement method may not. Added to this is the belief, held by some authorities, that in small rooms, reverberation time is a parameter of little or no value.

Understanding of small room acoustics, while being extensively researched over decades, is still very much in the dark ages. Main reason is that much of previous studies haven't been published and shared much in the open. It was reserved for organizations which required membership and it was long before the time of internet. Acousticians today mainly transfer their knowledge from large room acoustics to small rooms and it ends quite badly because they have completely different acoustic properties. It also a problem for large concert halls since there are small room acoustic challenges there to different degrees.
 

Bjorn

Major Contributor
Audio Company
Forum Donor
Joined
Feb 22, 2017
Messages
1,286
Likes
2,562
Location
Norway
In any case, as I mentioned above, this slide https://docs.google.com/presentatio...tsaowbpiFvjs_epoBvaTv28o7w/edit#slide=id.i165 suggests to me that the overall frequency spectra has not been unduly compromised by the use of overly thin absorption elements, despite your concerns.
You need to look at other measurements. Smoothed frequency response is only showing a small part. Specular reflections will not be treated evently with thin absorption. That we know for sure.
 

Kvalsvoll

Addicted to Fun and Learning
Audio Company
Joined
Apr 25, 2019
Messages
878
Likes
1,643
Location
Norway
You need to look at other measurements. Smoothed frequency response is only showing a small part. Specular reflections will not be treated evently with thin absorption. That we know for sure.

Exactly. The measurements in the presentation does not provide useful information about the acoustic properties of the room.

If perforated metal plates are used on walls, they may work as membrane absorbers at low frequencies, however, no mention of such here, only shows additional layer of panels inside the room, where flex of those will affect bass frequencies. Like in any other room. No wide-band absorption.

Frequency response in a small room says very little about acoustic properties. Generally, frequency response will be more smooth and even with more reflective energy, so it can actually be misleading.

You need to look at the decay profile across the frequency spectrum.
 

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
It's based on measurements and methods that are not applicable to small rooms as has been discussed in this thread.

It's pretty clear they don't understand much about small room acoustics and trying the best to find a middle road.
It was indeed discussed earlier in the thread but there wasn't much agreement with your pov. ;)
 
Last edited:

March Audio

Master Contributor
Audio Company
Joined
Mar 1, 2016
Messages
6,378
Likes
9,317
Location
Albany Western Australia
Exactly.

If perforated metal plates are used on walls, they may work as membrane absorbers at low frequencies, however, no mention of such here, only shows additional layer of panels inside the room, where flex of those will affect bass frequencies. Like in any other room. No wide-band absorption.

.
I confess I haven't read everything in the thread so I may be taking your reference to perforated metal plates out of context however it is a common technique in industrial environments (I have worked in noise modelling / mitigation in this area - Soundplan) to use perforated sheets over walls with a depth of fibreglass material / rockwool as an absorber. It works. Configuration/depth will obviously affect the specific response.
 
Last edited:

Kvalsvoll

Addicted to Fun and Learning
Audio Company
Joined
Apr 25, 2019
Messages
878
Likes
1,643
Location
Norway
I confess I haven't read everything in the thread so I may be taking your reference to perforated metal plates out of context however it is a common technique in industrial environments (I have worked in noise modelling in this area - Soundplan) to use perforated sheets over walls with a depth of fibreglass material / rockwool as an absorber. It works. Configuration/depth will obviously affect the response.

In post #193, talking about the construction of Harman listening rooms: .. "The inner walls and ceiling of the double-wall IAC shell are made of heavy gauge steel panels separated 10 cm and filled with fiberglass. The inner surfaces are perforated with 2.34-mm openings to provide substantial sound absorption inside the room.

@Bjorn had some comments on this, in post #195.
 
Top Bottom